Skip to contentSkip to conversation

iio: pressure: bmp280: fix bmp580 temp read

PATCH In review Adam Rizkalla 3 participants 6 messages Jun 2, 2024 – Jun 17, 2024 IIO

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iio: pressure: bmp280: fix bmp580 temp read

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,
 
Signature · 1 line
2.40.1
maintainer
#1 lore raw
Quoted 9 lines · On Sun, 2 Jun 2024 15:12:01 -0500 Adam Rizkalla <[email protected]> wrote:

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

Thanks! I was just about to email about this fix currently being lost in linux-next.

Rather than dance around this, I'm going to pull the later part of Vasielios' series off the togreg tree for now then reapply later once I have the fix in my upstream.

It's a shame that we need to do the maths in driver to keep within range. Maybe we can be cheeky and avoid doing the division?
I believe the IIO core formatting code should be fine with that as it already casts up to 64 bits to do the maths.

(s64)*raw_temp * 250);

bmp580_temp_coeffs = {1, 18}

Anyhow, I've dropped (for now) the following patches from my togreg tree
iio: pressure: bmp280: Generalize read_{temp,press,humid}() functions
iio: pressure: bmp280: Add SCALE, RAW values in channels and refactorize them
iio: pressure: bmp280: Add triggered buffer support

Jonathan

Quoted 35 lines
---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,
 
#2 lore raw
Quoted 27 lines · Jonathan Cameron
Quoted 9 lines · On Sun, 2 Jun 2024 15:12:01 -0500 Adam Rizkalla <[email protected]> wrote:

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

Thanks! I was just about to email about this fix currently being lost in linux-next.

Rather than dance around this, I'm going to pull the later part of Vasielios' series off the togreg tree for now then reapply later once I have the fix in my upstream.

It's a shame that we need to do the maths in driver to keep within range. Maybe we can be cheeky and avoid doing the division?
I believe the IIO core formatting code should be fine with that as it already casts up to 64 bits to do the maths.

(s64)*raw_temp * 250);

Quoted 9 lines

bmp580_temp_coeffs = {1, 18}

Anyhow, I've dropped (for now) the following patches from my togreg tree
iio: pressure: bmp280: Generalize read_{temp,press,humid}() functions
iio: pressure: bmp280: Add SCALE, RAW values in channels and refactorize them
iio: pressure: bmp280: Add triggered buffer support

Jonathan

Hi Jonathan, Adam,

I had mentioned it here [1], when I applied v7 of the patches that this would happen. No worries though, I already have the new version of these 3 patches that include Adam's fix, so when we have it upstream I can resubmit them (no need to keep a note Jonathan, I have it) and I think we will be fine!

Cheers,
Vasilis

[1]: https://lore.kernel.org/linux-iio/[email protected]/T/#t

Quoted 36 lines
Quoted 35 lines
---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,
 
maintainer
#3 lore raw
Quoted 38 lines · On Tue, 4 Jun 2024 21:19:04 +0200 Vasileios Amoiridis <[email protected]>…
Quoted 27 lines · Jonathan Cameron
Quoted 9 lines · On Sun, 2 Jun 2024 15:12:01 -0500 Adam Rizkalla <[email protected]> wrote:

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

Thanks! I was just about to email about this fix currently being lost in linux-next.

Rather than dance around this, I'm going to pull the later part of Vasielios' series off the togreg tree for now then reapply later once I have the fix in my upstream.

It's a shame that we need to do the maths in driver to keep within range. Maybe we can be cheeky and avoid doing the division?
I believe the IIO core formatting code should be fine with that as it already casts up to 64 bits to do the maths.

(s64)*raw_temp * 250);

Quoted 9 lines

bmp580_temp_coeffs = {1, 18}

Anyhow, I've dropped (for now) the following patches from my togreg tree
iio: pressure: bmp280: Generalize read_{temp,press,humid}() functions
iio: pressure: bmp280: Add SCALE, RAW values in channels and refactorize them
iio: pressure: bmp280: Add triggered buffer support

Jonathan

Quoted 6 lines

Hi Jonathan, Adam,

I had mentioned it here [1], when I applied v7 of the patches that this would happen. No worries though, I already have the new version of these 3 patches that include Adam's fix, so when we have it upstream I can resubmit them (no need to keep a note Jonathan, I have it) and I think we will be fine!

Excellent. The fix is no in the upstream of my togreg branch.

I've only pushed it out as testing for now, but if you send a version based on my testing branch or 20460472952 (currently char-misc-next head) that would be great.

Thanks,

Jonathan

Quoted 42 lines

Cheers,
Vasilis

[1]: https://lore.kernel.org/linux-iio/[email protected]/T/#t

Quoted 36 lines
Quoted 35 lines
---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,
 
#4 lore raw
Quoted 48 lines · Jonathan Cameron
Quoted 38 lines · On Tue, 4 Jun 2024 21:19:04 +0200 Vasileios Amoiridis <[email protected]>…
Quoted 27 lines · Jonathan Cameron
Quoted 9 lines · On Sun, 2 Jun 2024 15:12:01 -0500 Adam Rizkalla <[email protected]> wrote:

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

Thanks! I was just about to email about this fix currently being lost in linux-next.

Rather than dance around this, I'm going to pull the later part of Vasielios' series off the togreg tree for now then reapply later once I have the fix in my upstream.

It's a shame that we need to do the maths in driver to keep within range. Maybe we can be cheeky and avoid doing the division?
I believe the IIO core formatting code should be fine with that as it already casts up to 64 bits to do the maths.

(s64)*raw_temp * 250);

Quoted 9 lines

bmp580_temp_coeffs = {1, 18}

Anyhow, I've dropped (for now) the following patches from my togreg tree
iio: pressure: bmp280: Generalize read_{temp,press,humid}() functions
iio: pressure: bmp280: Add SCALE, RAW values in channels and refactorize them
iio: pressure: bmp280: Add triggered buffer support

Jonathan

Quoted 6 lines

Hi Jonathan, Adam,

I had mentioned it here [1], when I applied v7 of the patches that this would happen. No worries though, I already have the new version of these 3 patches that include Adam's fix, so when we have it upstream I can resubmit them (no need to keep a note Jonathan, I have it) and I think we will be fine!

Quoted 9 lines

Excellent. The fix is no in the upstream of my togreg branch.

I've only pushed it out as testing for now, but if you send a version based on my testing branch or 20460472952 (currently char-misc-next head) that would be great.

Thanks,

Jonathan

Hi Jonathan,

Thank for letting me know, I can send it by tomorrow.

Cheers,
Vasilis

Quoted 43 lines
Quoted 42 lines

Cheers,
Vasilis

[1]: https://lore.kernel.org/linux-iio/[email protected]/T/#t

Quoted 36 lines
Quoted 35 lines
---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,
 
#5 lore raw
Quoted 59 lines · Vasileios Amoiridis
Quoted 48 lines · Jonathan Cameron
Quoted 38 lines · On Tue, 4 Jun 2024 21:19:04 +0200 Vasileios Amoiridis <[email protected]>…
Quoted 27 lines · Jonathan Cameron
Quoted 9 lines · On Sun, 2 Jun 2024 15:12:01 -0500 Adam Rizkalla <[email protected]> wrote:

Fix overflow when storing bmp580 temperature reading and preserve sign.

This patch re-applies the fix in [1] after the merge conflict resolution mentioned in [2].

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/linux-kernel/[email protected]/

Thanks! I was just about to email about this fix currently being lost in linux-next.

Rather than dance around this, I'm going to pull the later part of Vasielios' series off the togreg tree for now then reapply later once I have the fix in my upstream.

It's a shame that we need to do the maths in driver to keep within range. Maybe we can be cheeky and avoid doing the division?
I believe the IIO core formatting code should be fine with that as it already casts up to 64 bits to do the maths.

(s64)*raw_temp * 250);

Quoted 9 lines

bmp580_temp_coeffs = {1, 18}

Anyhow, I've dropped (for now) the following patches from my togreg tree
iio: pressure: bmp280: Generalize read_{temp,press,humid}() functions
iio: pressure: bmp280: Add SCALE, RAW values in channels and refactorize them
iio: pressure: bmp280: Add triggered buffer support

Jonathan

Quoted 6 lines

Hi Jonathan, Adam,

I had mentioned it here [1], when I applied v7 of the patches that this would happen. No worries though, I already have the new version of these 3 patches that include Adam's fix, so when we have it upstream I can resubmit them (no need to keep a note Jonathan, I have it) and I think we will be fine!

Quoted 9 lines

Excellent. The fix is no in the upstream of my togreg branch.

I've only pushed it out as testing for now, but if you send a version based on my testing branch or 20460472952 (currently char-misc-next head) that would be great.

Thanks,

Jonathan

Quoted 7 lines

Hi Jonathan,

Thank for letting me know, I can send it by tomorrow.

Cheers,
Vasilis

Hi Jonathan,

Applied already, you can take the patch whenever you want, thanks for the help and the heads up!

Cheers,
Vasilis

Quoted 43 lines
Quoted 43 lines
Quoted 42 lines

Cheers,
Vasilis

[1]: https://lore.kernel.org/linux-iio/[email protected]/T/#t

Quoted 36 lines
Quoted 35 lines
---
 drivers/iio/pressure/bmp280-core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/iio/pressure/bmp280-core.c +4 -2
index 95c88b0e1c49..3a003843c79c 100644
@@ -1752,6 +1752,8 @@ static int bmp580_read_temp(struct bmp280_data *data, s32 *raw_temp)
 		dev_err(data->dev, "reading temperature skipped\n");
 		return -EIO;
 	}
+	*raw_temp = sign_extend32(*raw_temp, 23);
+	*raw_temp = ((s64)*raw_temp * 1000) / (1 << 16);
 
 	return 0;
 }
@@ -2154,7 +2156,7 @@ static irqreturn_t bmp580_buffer_handler(int irq, void *p)
 
 static const int bmp580_oversampling_avail[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
 static const u8 bmp580_chip_ids[] = { BMP580_CHIP_ID, BMP580_CHIP_ID_ALT };
-static const int bmp580_temp_coeffs[] = { 1000, 16 };
+static const int bmp580_temp_coeffs[] = { 1, 0 };
 static const int bmp580_press_coeffs[] = { 1, 64000};
 
 const struct bmp280_chip_info bmp580_chip_info = {
@@ -2184,7 +2186,7 @@ const struct bmp280_chip_info bmp580_chip_info = {
 	.iir_filter_coeff_default = 2,
 
 	.temp_coeffs = bmp580_temp_coeffs,
-	.temp_coeffs_type = IIO_VAL_FRACTIONAL_LOG2,
+	.temp_coeffs_type = IIO_VAL_INT,
 	.press_coeffs = bmp580_press_coeffs,
 	.press_coeffs_type = IIO_VAL_FRACTIONAL,